Chemistry Reference
In-Depth Information
of the binding fragments showed an inhibition of
>
50% at 250 M concentration and an
additional five fragments showed
>
50% inhibition at 500 M. All the binding fragments
showed at least some inhibition at 500 M, with the exception of one that appeared to be
an activator. Dose-response curves were collected for the most potent inhibitor fragments
and the two most potent fragment inhibitors showed remarkable 0.8 and 13 M potencies.
This translated to ligand efficiencies of 0.55 and 0.51 kcal mol
−
1
, respectively, indicating
very favorable binding interactions. The most potent fragment hits were shown to compete
for binding with the knownATP-pocket binder staurosporine. Despite considerable efforts,
no crystal structure of this target kinase was obtained.
In order to check if it was possible to optimize the binding scaffolds, close fragment
analogues to the two most potent fragment hits were tested in the activity assay. Several
of these analogues showed inhibition but none was more potent than the original fragment
hits, which was hardly surprising considering their high ligand efficiency. The in-house
compound collection and commercial sources were then searched for larger compounds in
which the most potent fragment hits are present as a substructure (exact or very similar to
the fragment hit). Seven and 12 compounds were found for the 0.8 and 13 M fragments,
respectively. Among the seven analogues for the 0.8 M fragment, one was active (IC
50
=
4 M; LE
0.28 kcal mol
−
1
), but this compound was fairly large and would probably
require cumbersome chemistry to develop, so this track was dropped. The substructure
analogues of the 13 M fragment hit looked more promising. Of the 12 analogues tested,
three showed good potency (IC
50
=
=
0.33, 0.35 and 0.34 kcal mol
−
1
,
respectively) and competed for binding with staurosporine, as seen by STD (see Figure 4.8).
These three compounds were diverse, highly soluble and would constitute the first com-
pounds in three series. Especially the 1 M analogue was considered very interesting since
there was no mentioning in the literature of this type of compound. Further, the compound
showed significant inhibition (
>
50% inhibition at 10 M) to only one other kinase in a
selectivity panel consisting of 30 diverse kinases, which compared favorably with nearly
all of the compounds found by the HTS campaign. In conclusion, fragment screening
coupled with a standard biochemical kinase assay succeeded in identifying several lead
series, including a truly novel lead series, to a Ser/Thr kinase target despite the fact that
no structural information and no synthetic chemistry resources were used. Only a small
number of in-house and commercially available analogues had been tested up to this stage.
This gave a significant boost to the outcome of the HTS from which three promising hit
series had been identified, all from kinase targeted libraries.
1, 2 and 6 M; LE
=
4.6.3 Difficult Targets for Fragment-based Screening by NMR
As discussed above, fragment-based screening does in general succeed in finding good
lead series with high solubility and ligand efficiency as well as novel binding scaffolds.
However, there are types of targets with associated difficulties.
Target proteins forming large multimeres.
It is important to ascertain that the target pro-
tein is monodisperse at the protein concentration and in the buffer that is going to be used
for the screening. In one case with a large protease target, this was not properly checked
before the fragment screen. STD was used to screen fragments at a protein and fragment
concentration of 1 and 100 M, respectively. The hit rate turned out to be extremely high,
